Vulnerability in vLLM (CVE-2026-100654)
Summary
Vulnerability in vLLM (CVE-2026-100654). Risk of unauthorized operations or information disclosure. Exploitable via POST /v1/completions.
Original description
The description as published by the source.
vLLM before 0.29.0 accepts user-controlled stop_token_ids on the OpenAI-compatible POST /v1/completions and POST /v1/chat/completions endpoints but validates only that the values are integers, not that each token id is within the model vocabulary/logits range. When min_tokens > 0, the stop token ids are used as logits indices to suppress stop tokens, so an out-of-range id reaches a CUDA indexing operation (index_put_) and triggers a device-side assertion. An authenticated API user can send a single malformed completion request that returns 500 Internal Server Error and puts EngineCore into a fatal state, causing subsequent requests to fail until the service is restarted (denial of service).
Source: NIST National Vulnerability Database NVD GitHub Advisory
Auto-generated summary
Assembled by SecPulse from the source data (CVSS, CWE, affected versions) using templates — no AI involved. Check the original description and references below for exact details.
A vulnerability tracked as CVE-2026-100654 has been found in vLLM.
Attackers can target a specific entry point like POST /v1/completions over the network to misuse the product.
Risk of unauthorized operations or information disclosure. CVSS score: 6.5/10.
What to do: apply the vendor's official patched release.
If unsure, ask your IT team or search "vLLM CVE-2026-100654" on the vendor's site.
CVE-2026-100654 (vLLM) — CWE-129 / CVSS v3 6.5
Attack vector: remote (network-reachable) / no user interaction
Attack surface: POST /v1/completions / POST /v1/chat/completions
Plan: 1) Audit SBOM/dependencies, 2) Stage→prod upgrade, 3) Add WAF/proxy monitoring on affected endpoints, 4) Hunt IOCs in logs.
Refs: see the GHSA / vendor advisory / patched release linked on this page.
A vulnerability (CWE-129) exists in vLLM. Attackers reach the vulnerable code path via POST /v1/completions.
Check the vendor advisory for the affected versions of vLLM. Attack surface: POST /v1/completions / POST /v1/chat/completions.
Severity: Medium (CVSS 6.5/10). Risk of unauthorized operations or information disclosure.
Update to the patched release per vendor advisory.
Until the patch is applied: disable the affected feature, apply WAF rules, or restrict access via network ACLs.
Search webserver/proxy logs for unusual POST /v1/completions requests with malformed payloads or SQL meta-characters. Run grep -r 'vLLM' . against your dependency files (package-lock.json, requirements.txt, go.sum) to find affected services.
Response actions
Concrete steps and command examples for SOC/SRE teams to execute in order
-
1Identify exposure identify
grep -r 'vLLM' . | grep -v node_modulesリポジトリと本番環境の依存ファイル (package-lock.json / requirements.txt / go.sum / Gemfile.lock 等) で
vLLMを grep し、稼働しているサービス・バージョンを把握する。 -
2Hunt for indicators of compromise detect
grep 'POST /v1/completions' /var/log/nginx/access.log | grep -E '(unusual_payload|sqli_pattern)'アクセスログで
POST /v1/completionsへの異常なリクエスト (不正な認証ヘッダ・SQLメタ文字)を過去 30〜90日分捜索。WAF/SIEM があれば該当パスのアラート発火履歴を確認。 -
3Post-deployment verification verify
Replay attack against POST /v1/completions on staging to confirm patch closes the vectorパッチ適用後、ステージングで PoC または同等の悪用パターンを再現して脆弱性が閉じたことを確認。本番では Step 2 と同じログクエリでアラート再発が無いか継続監視。
References
- advisory disclosure@vulncheck.com
- patch disclosure@vulncheck.com
- web nvd.nist.gov
- web github.com